R600/SI: Add pattern for fp_to_uint

This fixes the F2U opcode for the Mesa driver.

Patch by: Marek Olšák

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188418 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/R600/SIInstructions.td b/lib/Target/R600/SIInstructions.td
index dc146091df6f9669573c5e0a66a1b2de3e847d64..7393e1de563832e1b9559894ef03de51ed1a46c2 100644 (file)
--- a/lib/Target/R600/SIInstructions.td
+++ b/lib/Target/R600/SIInstructions.td
@@ -615,7 +615,9 @@ defm V_CVT_F32_I32 : VOP1_32 <0x00000005, "V_CVT_F32_I32",
 defm V_CVT_F32_U32 : VOP1_32 <0x00000006, "V_CVT_F32_U32",
   [(set f32:$dst, (uint_to_fp i32:$src0))]
 >;
-defm V_CVT_U32_F32 : VOP1_32 <0x00000007, "V_CVT_U32_F32", []>;
+defm V_CVT_U32_F32 : VOP1_32 <0x00000007, "V_CVT_U32_F32",
+  [(set i32:$dst, (fp_to_uint f32:$src0))]
+>;
 defm V_CVT_I32_F32 : VOP1_32 <0x00000008, "V_CVT_I32_F32",
   [(set i32:$dst, (fp_to_sint f32:$src0))]
 >;

diff --git a/test/CodeGen/R600/fp_to_uint.ll b/test/CodeGen/R600/fp_to_uint.ll
index 2a365f9573a9ab7347813562dcc0bdd666e02e04..0d07a616aa12aa0fe0c5ed1713e77fd9878906d8 100644 (file)
--- a/test/CodeGen/R600/fp_to_uint.ll
+++ b/test/CodeGen/R600/fp_to_uint.ll
@@ -1,8 +1,12 @@
-; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
+; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s --check-prefix=R600-CHECK
+; RUN: llc < %s -march=r600 -mcpu=SI | FileCheck %s --check-prefix=SI-CHECK
 
-; CHECK: @fp_to_uint_v2i32
-; CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
-; CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
+; R600-CHECK: @fp_to_uint_v2i32
+; R600-CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
+; R600-CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
+; SI-CHECK: @fp_to_uint_v2i32
+; SI-CHECK: V_CVT_U32_F32_e32
+; SI-CHECK: V_CVT_U32_F32_e32
 
 define void @fp_to_uint_v2i32(<2 x i32> addrspace(1)* %out, <2 x float> %in) {
   %result = fptoui <2 x float> %in to <2 x i32>
@@ -10,11 +14,16 @@ define void @fp_to_uint_v2i32(<2 x i32> addrspace(1)* %out, <2 x float> %in) {
   ret void
 }
 
-; CHECK: @fp_to_uint_v4i32
-; CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
-; CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
-; CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
-; CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
+; R600-CHECK: @fp_to_uint_v4i32
+; R600-CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
+; R600-CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
+; R600-CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
+; R600-CHECK: FLT_TO_UINT * T{{[0-9]+\.[XYZW], PV\.[XYZW]}}
+; SI-CHECK: @fp_to_uint_v4i32
+; SI-CHECK: V_CVT_U32_F32_e32
+; SI-CHECK: V_CVT_U32_F32_e32
+; SI-CHECK: V_CVT_U32_F32_e32
+; SI-CHECK: V_CVT_U32_F32_e32
 
 define void @fp_to_uint_v4i32(<4 x i32> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
   %value = load <4 x float> addrspace(1) * %in